Brain-machine interface (BMI) with user interface (UI) aware controller

Abstract

Methods involving interpreting signals from a brain-machine interface (BMI) are described, as well as methods involving adjusting an implanted or wearable BMI device. The method includes receiving neural signals from a brain of a subject into a BMI decoder. The method includes determining an activity change of the subject based on a sensor. The method includes routing the neural signals from a first model to a second model in the BMI decoder based on the determined activity change. The method includes translating, using the second model in the BMI decoder, the neural signals into a command. The method includes sending the command to a controller.

What is claimed is: 
     
       1. A method of interpreting signals from a brain-machine interface (BMI), the method comprising:
 receiving a first set of neural signals from a brain of a subject into a BMI decoder; 
 translating, using a first model in the BMI decoder, the first set of neural signals into a first command; 
 detecting, from further neural signals from the subject, a frustration from the subject; 
 inhibiting the first command by sending a cancelation of the first command, the inhibiting based on the detecting; 
 routing a second set of neural signals from the brain of the subject to a second model in the BMI decoder based on the detecting; 
 interpreting, using the second model, the second set of neural signals into a second command; and 
 sending the second command. 
 
     
     
       2. The method of  claim 1  wherein the second command is sent to a cursor, a keyboard, a robotic arm, or a wheelchair. 
     
     
       3. The method of  claim 1  wherein the neural signals pass through metal electrodes in a cerebral cortex of the brain. 
     
     
       4. The method of  claim 3  further comprising:
 converting, through an analog-to-digital converter (ADC), voltages or currents from the electrodes; 
 detecting spikes from the voltages or currents; and 
 forwarding the spikes as the neural signals. 
 
     
     
       5. The method of  claim 1  wherein the first or second model includes binning neural spikes as a function of frequency. 
     
     
       6. A method of adjusting an implanted or wearable brain-machine interface (BMI) device, the method comprising:
 receiving neural signals from a brain of a subject into an implanted or wearable BMI device; 
 determining an activity change of the subject based on a sensor; 
 switching from a first compression algorithm to a second compression algorithm based on the determined activity change; 
 compressing, using the second compression algorithm in the BMI device, the neural signals into a data stream; 
 sending the data stream to a BMI controller off-board the subject. 
 
     
     
       7. The method of  claim 6  wherein the activity change is a change between the subject interfacing with a graphical user interface (GUI) to the subject manipulating a physical device. 
     
     
       8. The method of  claim 7  wherein the activity change is from the subject interfacing with a graphical user interface (GUI) to the subject manipulating a physical device. 
     
     
       9. The method of  claim 7  wherein the interfacing with the GUI involves moving a cursor, entering text, or selecting words, and the manipulating a physical device involves operating a robotic arm or steering a wheelchair. 
     
     
       10. The method of  claim 6  wherein the neural signals pass through metal electrodes in a cerebral cortex of the brain. 
     
     
       11. The method of  claim 10  further comprising:
 converting, through an analog-to-digital converter (ADC), voltages or currents from the electrodes; 
 detecting spikes from the voltages or currents; and 
 forwarding the spikes as the neural signals. 
 
     
     
       12. The method of  claim 6  wherein the first or second compression algorithm includes binning neural spikes as a function of frequency.